The development of the optical imaging techniques has significantly advanced the research on neuroscience, especially the use of fluorescent indicators and scanning microscopy. One-photon excitation (1PE) microscopy utilizes single-photon absorption where a fluorophore absorbs a single photon of light, usually in the visible range, and provides a simple and versatile imaging method with approachable light sources. Since 1990, when Denk pioneered two-photon laser scanning fluorescence microscopy [1], the use of non-linear imaging techniques has grown exponentially. Two-photon excitation (2PE) and three-photon excitation (3PE) microscopy are advantageous imaging modalities when used in conjunction with genetically encoded calcium indicators (GECIs) for calcium imaging because the near infrared wavelength has less scattering and absorption and thus a greater penetration depth in the biological tissue than the visible wavelength. In comparison to 1PE excitation, multi-photon excitation exploits the advantages of optical nonlinearity by having the fluorophores simultaneously absorb several long-wavelength photons. This significantly improves the signal-to-background ratio, axial resolution and meanwhile reduces the phototoxicity, and has revolutionized the ability to visualize and understand the intricate workings of neuronal networks, enabling a spatially high-resolution imaging of dynamic biological processes. The selection of 1PE or multiphoton excitation for functional imaging depends on the specific application, depth of imaging, sample properties (i.e., transparent or vague scattering one) and the photodamage considerations. Generally speaking, the 1PE is simpler and more cost-effective, and 2PE systems are more complex and expensive, requiring specialized femtosecond lasers and precise alignment. For shallower imaging depth and sample with higher tolerate level of photodamage, 1PE is generally sufficient. For applications with a high requirement on the SNR and in vivo capabilities, 2PE is often the method of choice. However, conventional 1PE microscopes do not offer depth resolution, which makes other planes the background of the target plane and decreases the signal to background ratio (SBR). To overcome this problem, we aim to build the depth-sensitive PSF to reconstruct the information from each depth from a single camera image. For 2PE microscopy, to achieve a high SNR, usually they come with beam steering function from scanning mirrors. However, conventional point-by-point scanning microscopy has a trade-off between sampling speed and spatiotemporal resolution and is typically susceptible to oversampling, and the conventional volumetric data collection requires the slowly moved mechanical components, when imaging a large field-of-view of sample or the volumetric sample, the temporal resolution becomes very poor. This has severely limited its applications in calcium imaging where it is important to record the fluorescence in high temporal resolution in order to capture the neuronal signals faithfully. In addition, the phototoxicity and photobleaching from the raster-scanning schematics makes the traditional methods suffer from limited penetration depth, which can impede the efficiency of neuronal imaging.In response to the challenge in the low frame rate from image acquisition, this dissertation focuses on the development of innovative illumination techniques under the context of scanning microscopy and volumetric collection methods where no scanning is necessarily required. The primary objectives of this dissertation are twofold in providing the high-throughput imaging.First, to introduce and validate a novel illumination technique that leverages adaptive illumination to achieve high-speed imaging with a significantly reduced phototoxicity. In this dissertation, an innovative adaptive sampling strategy in multiphoton microscopy is proposed, which can significantly boost the imaging speed while maintaining a high imaging quality and low laser power. The microscope uses a new adaptive sampling scheme with line scanning, which also has the capability to switch to conventional point scanning, and it has potential to be applied to the block-by-block scanning to significantly increase the scanning speed in the future. Instead of building images pixel-by-pixel via scanning a diffraction-limited spot across the sample, this scheme only illuminates the regions of interest (i.e., neuronal cell bodies), and samples a large area of them in a single measurement. Such a scheme significantly increases the imaging speed and reduces the overall laser power on the brain tissue. Using this approach, the high-speed imaging of the neuronal activity in mouse cortex in vivo is performed. The method in this dissertation provides a new sampling strategy in laser-scanning two-photon microscopy and will be powerful for high-throughput imaging of neural activity. Secondly, the primary objective is to develop a novel volumetric collection method designed for 1PE microscopy that integrates multiple light sheet microscopy with advanced computational algorithms for fast 3D reconstruction on volume. In this dissertation, a novel V-shaped PSF approach is proposed to significantly enlarge the depth of field (DOF) involving replacing the conventional tube lens with a pair of axicon lenses positioned behind the collection objective of the light sheet microscope. This configuration allows for the capture of comprehensive three-dimensional (3D) information from a single-shot image, which can then be reconstructed using a deep neural network. Through these innovations, this work seeks to push the boundaries of optical imaging in neuroscience, offering new insights into neuronal activity and connectivity.By improving imaging speed and reducing the overall laser power on the brain tissue, these advancements aim to enhance the clarity and depth of neuronal imaging, thereby facilitating more comprehensive studies of brain function and pathology.